The present invention relates to the field of image sensors and display devices.
Microlenses have long been used in imaging devices to focus light on sensors including charge couple device (CCD) sensors and complementary metal oxide semiconductor (CMOS) sensors. The microlenses significantly improve the light sensitivity of the imaging device by collecting light from a large light collecting area and focusing it on a small light sensitive area of the sensor. The ratio of the overall light collecting area of a sensor to the light sensitive area of the sensor is defined to be a fill factor. Typical fill factors in prior art designs are less than 50%.
One prior art method of generating a color image signal is shown in FIG. 1A. Light from a subject to be imaged comes in as light rays 104 and passes through a set of microlenses 108, 112, 116. The microlenses are formed on a planarization layer 120. After passing through the planarization layer 120, the light 104 is filtered by color filters 124, 128, 132 which together form a color filter array. Each color filter 124, 128, 132 in the color filter array only allows light of a specific color to pass through. A xe2x80x9ccolorxe2x80x9d is defined to be light having a specific range of frequencies. Typical color filters 124, 128, 132 used in the color filter array are red, green and blue filters (RGB) or cyan, magenta and yellow (CMY) filters. Each microlens and color filter combination corresponds to a sensor 136, 140, 144. Each sensor is a light sensitive device capable of converting the intensity of light striking the sensor 136, 140, 144 into an electrical signal. A microlens, color filter, and sensors such as sensors 136, 140, 144 correspond to a pixel of an image. The sensors 136, 140, 144 are in close proximity to each other, and each sensor receives filtered light from a corresponding color filter 124, 128, 132. By combining the output of the sensors 136, 140, 144, a processor, such as a graphics processor, can determine the approximate intensity and color of light striking the area in the proximity of sensor 136, 140, 144. By creating an array of such sensors (red sensor 160, blue sensor 164, green sensor 168) as shown in FIG. 1B, an overall color image can be reconstructed.
The fabrication of separate microlenses, color filters, and image sensors in the structure illustrated in FIGS. 1A and 1B has several disadvantages. For example, one disadvantage of the traditional structure is that many process steps are needed to form a first layer 148 including the sensors 136, 140, 144; a second layer 152 including the color filters 124, 128, 132, and a third planarization layer 156 to support microlenses 108, 112, 116.
Another disadvantage of the current structure is that the microlenses 108, 112, 116 are separated from the corresponding image sensors 136, 140, 144 by the planarization layer 156 and the color filter layer 152. The separation reduces the light reaching the sensors 136, 140, 144 because some light is absorbed passing through the multiple layers 152, 156. Furthermore, the separation results in increased crosstalk between pixels. xe2x80x9cCrosstalkxe2x80x9d results when off axis light strikes a microlens such as microlens 112 at an obtuse angle of incidence. The off-axis light passes through planarization layers 156 and a color filter 128 missing the sensor 140 which corresponds to the color filter 128 and instead striking an adjacent sensor 136. Alternately, the off-axis light coming in through microlens 112 may pass between filters 124 and 128 and reach adjacent sensor 136 resulting in an erroneous readings from the image sensor 136.
Additional disadvantages of the currect micro-lens filter combinations include the additional process steps being used to fabricate the multi-level structure of FIG. 1, the decreased reliability resulting from separation of layers 148, 152, 156 and the increased material costs used to fabricate separate transparent microlenses 108, 112, 116, color filters 124, 128, 132, and associated planarization layer 156.
A second use of the microlens, color filter layer, structure is in color display devices. FIG. 2 illustrates an example of using the microlens color filter structure in a thin film transfer (TFT) liquid crystal display device. In FIG. 2, light from a backlight or other light source 204 passes through a color filter layer 208 containing color filters 212, 216 and 220. The color filters 212, 216, 220 are typically different colors allowing only one color of light to pass through each filter. Microlenses 224, 228 and 232 in microlens layer 236 focuses the light from corresponding color filters 212, 216, 220 through a substrate 240 and a liquid crystal display (LCD) layer 244 to a TFT substrate 248. Each TFT switch 252, 256, 260 corresponds to a corresponding color filter 212, 216, 220. By controlling the amount of light passing through each switch 252, 256, 260, the output of each color filter 212, 216, 220 can be controlled. Combining the outputs of the color filters and TFT switches generates the output of a pixel of the color display device.
Display devices formed using the described techniques suffer from the previously described disadvantages including (1) difficulty in fabrication; (2) crosstalk between filters and switches caused by the increased separation generated by the microlens layer; and (3) problems with device reliability resulting from adhesion between multiple layers and increased material costs resulting from the necessity for multiple layers.
Thus an improved method for forming microlens and color filter structures is desired.
The present invention describes a method of forming a color microlens array on a semiconductor substrate. The method involves depositing a colored microlens resist on a semiconductor surface. The colored microlens resist is patterned and then baked to cause flowing of the colored microlens resist resulting in a color microlens with a curved surface.